Update llama.cpp API 20260106

Author: JamePeng

llama_cpp/_internals.py

@@ -92,6 +92,15 @@ def n_cls_out(self) -> int:
     def n_embd(self) -> int:
         return llama_cpp.llama_model_n_embd(self.model)
 
+    def n_embd_inp(self) -> int:
+        return llama_cpp.llama_model_n_embd_inp(self.model)
+
+    def n_embd_out(self) -> int:
+        return llama_cpp.llama_model_n_embd_out(self.model)
+
+    def n_layer(self) -> int:
+        return llama_cpp.llama_model_n_layer(self.model)
+
     def n_head(self) -> int:
         return llama_cpp.llama_model_n_head(self.model)
 

llama_cpp/llama.py

@@ -2351,14 +2351,42 @@ def n_ctx(self) -> int:
         """Return the context window size."""
         return self._ctx.n_ctx()
 
+    def n_ctx_train(self) -> int:
+        """Return the training context window size."""
+        return self._model.n_ctx_train()
+
     def n_embd(self) -> int:
         """Return the embedding size."""
         return self._model.n_embd()
 
+    def n_embd_inp(self) -> int:
+        """Return the input embedding size."""
+        return self._model.n_embd_inp()
+
+    def n_embd_out(self) -> int:
+        """Return the output embedding size."""
+        return self._model.n_embd_out()
+
+    def n_layer(self) -> int:
+        """Return the n_layer value."""
+        return self._model.n_layer()
+
+    def n_head(self) -> int:
+        """Return the head size."""
+        return self._model.n_head()
+
     def n_head_kv(self) -> int:
         """Return the head_kv size."""
         return self._model.n_head_kv()
 
+    def n_swa(self) -> int:
+        """Return the swa size."""
+        return self._model.n_swa()
+
+    def n_params(self) -> int:
+        """Returns the total number of parameters in the model"""
+        return self._model.n_params()
+
     def n_vocab(self) -> int:
         """Return the vocabulary size."""
         return self._model.n_vocab()

llama_cpp/llama_cpp.py

@@ -1562,6 +1562,12 @@ def llama_model_n_embd_inp(model: llama_model_p, /) -> int:
     ...
 
 
+# LLAMA_API int32_t llama_model_n_embd_out (const struct llama_model * model);
+@ctypes_function("llama_model_n_embd_out", [llama_model_p_ctypes], ctypes.c_int32)
+def llama_model_n_embd_out(model: llama_model_p, /) -> int:
+    ...
+
+
 # LLAMA_API int32_t llama_model_n_layer    (const struct llama_model * model);
 @ctypes_function("llama_model_n_layer", [llama_model_p_ctypes], ctypes.c_int32)
 def llama_model_n_layer(model: llama_model_p, /) -> int: